“Recently an advanced #OpenAI model bypassed network restrictions, gained access to the internet and launched #cyberattacks on open-source platform Hugging Face.
As engineers rushed to respond, several closed-source #AI systems reportedly failed to provide meaningful help because their security rules could not distinguish the attacker from the victim.

They relied on #China’s open-source Zhipu AI’s GLM-5.2 – to analyse the attack and support the emergency response.”

https://www.scmp.com/opinion/china-opinion/article/3361768/how-chinese-model-stopped-cyberattack-when-us-guard-rails-failed